EQ fico 08 vs EQ fico bankcard 08

[Anonymous]

If you are curious when your EQ FICO 8 Classic hits 850, you are doing the right thing by joining Langley (or pulling the true Classic score some other way).  Your idea about subtracting 30 (or any other number) won't work.  There's no way to infer the precise Classic score from the BE, though of course you could infer with fair confidence when it was > 840.  Even if you had a BE of 900 you could not infer with certainty a Classic of 850, though it would be a good bet.

Congrats on having such good scores!  My TU hit 850 back in Feb/March, which was kinda fun.

[Anonymous]

Here's how my EQ Classic 8 vs BE 8 compared over the last 6 cycles, generally using the same bureau data:

Classic 8       BE 8     Variance

847                  871          24

825                  845          20

828                  848          20

831                  853          22

828                  851          23

831                  850          19

The variance overall seems pretty consistent. 

[sarge12]

---

@Anonymous wrote:

Here's how my EQ Classic 8 vs BE 8 compared over the last 6 cycles, generally using the same bureau data:

Classic 8       BE 8     Variance

847                  871          24

825                  845          20

828                  848          20

831                  853          22

828                  851          23

831                  850          19

The variance overall seems pretty consistent. 

---

Good data BBS...thanks!

[Thomas_Thumb]

---

@Anonymous wrote:

If you are curious when your EQ FICO 8 Classic hits 850, you are doing the right thing by joining Langley (or pulling the true Classic score some other way).  Your idea about subtracting 30 (or any other number) won't work.  There's no way to infer the precise Classic score from the BE, though of course you could infer with fair confidence when it was > 840.  Even if you had a BE of 900 you could not infer with certainty a Classic of 850, though it would be a good bet.

Congrats on having such good scores!  My TU hit 850 back in Feb/March, which was kinda fun.

---

The Fico 8 Bankcard algorithm looks has more score cards and weighs factors somewhat differently than Classic Fico 8. So, as CGID mentions, no robust conversion exists. However, a BC Fico 8 score of 900 is more difficult to achieve than the Classic Fico 8 850 score counterpart. My guess is a given profile must meet the Classic 850 metrics before a BC 900 can be realized. The below paste is relates to both Bankcard and Auto enhanced.

Here are the article links:

https://www.equifax.com/pdfs/corp/EFS-913-ADV-BEACON-09-Auto-Industry-Option.pdf

https://www.equifax.com/pdfs/corp/EFS-914-ADV-BEACON-09_BankcardIndustryOption.pdf

[Anonymous]

TT, what do you believe the factors are that impact the variance between an 850 Classic score and a 900 BCE score?  Since you've been able to top out both models you probably have a pretty good grasp of the top-end differences.

[Thomas_Thumb]

BBS,

Bankcard enhanced Fico 8 is more sensitive to # cards reporting than its Classic counterpart. The Auto enhanced counterpart looks specifically for an auto loan. I don't have an auto loan, open or closed, on file and suspect that' may be a reason why TU and EX Auto scores top out at 897 - while the BCEs versions were able to reach 900.

It would be helpful to know how the additional enhanced scorecard(s) segment profiles but, I have not found info detailing the segmentation.

[Anonymous]

What about when AZEO is in place?  If you look back up at Post 12 where I showed a 24 point variance between my EQ Classic and EQ BCE scores, both of those are with AZEO in place.  There's got to be other factors outside of number of cards with balances that BCE is more sensitive to, as I'm in an ideal place with respect to number of cards with balances (AZEO) and am only at 871. 

Is the BCE model possibly more critical of age of accounts factors?  All of my age of accounts factors are basically borderline in terms of being top-notch verses just below; I'd imagine in 6-12 months I'd be "ideal" across all of them.

Does the BCE model like to see a non-mortgage installment loan present?  I know such a loan (say, auto) in addition to a mortgage could be worth 5-10 points on the Classic model... perhaps on the BCE model it would be "worth" more?

[Thomas_Thumb]

There are undoubtedly many other differences. A couple that come to mind are:

1) Recent credit (including HPs for CLIs). I'd define recent in this case as 24 months age.

2) The enhanced versions looking for more accounts in general and more types of accounts specifically.

I have not had an inquiry on EX in more than 6 years. In August 2016 I had "no recent activity on revolving accounts". Auto score dropped substantially. In February 2016 I had 5 of 5 cards (6 of 6 including AU) reporting balances.

Fico Model Name	Mar-15	Jul-15	Nov-15	Feb-16	Mar-16	Jul-16	Aug-16	Oct-16	Mar-17	Apr-18
EX Bankcard Fico 8	898	898	900	895	898	900	892	897	900	900
EX Classic Fico 8	850	850	850	850	850	850	850	850	850	850
EX score difference	48	48	50	45	48	50	42	47	50	50

Fico Model Name	Mar-15	Jul-15	Nov-15	Feb-16	Mar-16	Jul-16	Aug-16	Oct-16	Mar-17	Apr-18
EX Bankcard Fico 8	898	898	900	895	898	900	892	897	900	900
EX Auto Fico 8	889	889	884	883	884	894	869	891	897	895
EX Score difference	9	9	16	12	14	6	23	6	3	5

[Anonymous]

You must have had a decent buffer than on the BCE model similar to your Classic 8 buffer, then. 

If you were able to allow all of your cards to report balances and only saw your 900 score go to 895, I'd say you had a good buffer in place on a model that as you stated is sensitive to number of cards with balances reported.

[Thomas_Thumb]

---

@Anonymous wrote:

You must have had a decent buffer than on the BCE model similar to your Classic 8 buffer, then. 

If you were able to allow all of your cards to report balances and only saw your 900 score go to 895, I'd say you had a good buffer in place on a model that as you stated is sensitive to number of cards with balances reported. [Yes, BCE is rather sensitive to # (or %) of cards reporting but less so with EX].

---

Overall, I don't see much of a top end buffer on BCE. Not sure how the additional 50 BCE points are allocated. Otherwise I would expect 900 to hold the way 850 does with Classic Fico 8.

EX puts less weight on # cards reporting than does either TU and EQ. See below comparison

CRA	Nov-15	Feb-16	Mar-16	Max-Min
Ag UT	1%	3%	2%	***
Cards	2 of 6	5 of 6	3 of 6	***
EQ BCE	886	874	882	12
TU BCE	900	880	899	20
EX BCE	900	895	898	5

Fico has 5 primary scoring factors with multiple attributes making up each category. Given each factor is allocated a certain number of points (or % of total points available for a given scorecard), a buffer for each is possible. That would help explain why some profiles with similar scores that appear to be on the same scorecard are more stable than others to changes in a specific attribute.

For a couple briefs on industry enhanced scores check out the Equifax links in the above post:

https://ficoforums.myfico.com/t5/Understanding-FICO-Scoring/EQ-fico-08-vs-EQ-fico-bankcard-08/m-p/5269900#M140550